Ｓ変換を前処理にしたニューラルネット風力発電出力予測

カテゴリ: 部門大会

論文No: 4

グループ名: 【B】平成29年電気学会電力・エネルギー部門大会

発行日: 2017/09/05

タイトル(英語): Application of S-Transform-based ANN to Wind Power Generation Output Forecasting

著者名: 大蔵 惣一朗(明治大学),森 啓之(明治大学)

著者名(英語): Okura Soichiro|Mori Hiroyuki

キーワード: 風力発電|ニューラルネットワーク|一般化ラジアル基底関数ネットワーク|S変換|進化的粒子群最適化二段階予測,wind power,artificial neural network,geberalized radial basis function network,S-Transform,EPSOTwo-Staged forecasting

要約(日本語): In this paper, a new ANN-based methods is proposed for wind power generation output forecasting. The proposed method makes use of GRBFN (Generalized Radial Basis Function Network) that is an extension of RBFN (Radial Basis Function Network). To improve the performance of GRBFN, this paper proposes three strategies: S-Transform of feature extraction, EPSO of evolutionary computation, and Two-Staged forecasting. S-Transform is used as a new feature extraction technique for input variables of GRBFN. EPSO is employed to evaluate the globally optimal weights between neurons to tune up GRBFN efficiently. Two-staged forecasting enhances the predicted value by constructing the error model. The proposed method is successfully applied to real data of wind power generation output.

PDFファイルサイズ: 602 Kバイト